Every few months, someone hands you the same argument: ADHD diagnoses have exploded, therefore doctors are handing them out too easily, therefore yours probably doesn’t “count” the way a childhood diagnosis would have. A study published September 9, 2026 in JAMA Psychiatry just took that argument apart with the most direct test anyone’s run on it yet. 2.1 million Danish kids, ten years, and one very specific question: do the people getting diagnosed today actually look different from the people who would’ve been diagnosed a decade ago?

They do. Just not in the direction the “overdiagnosis” crowd assumes.

TL;DR: What the Study Found

Question	Answer
The study	Changes in Characteristics Associated With ADHD and ASD Diagnoses Over Time, JAMA Psychiatry, Sep 9, 2026
Who ran it	Magnus Elias Tarp (Aarhus University) and Oleguer Plana-Ripoll (ISGlobal / Aarhus University), senior author
The data	2.1M+ Danish children and adolescents, 2012-2022; 71,317 diagnosed with ADHD or ASD vs. 713,170 undiagnosed peers
Low birth weight “advantage,” 2012	54% higher likelihood of diagnosis
Low birth weight “advantage,” 2022	17% higher likelihood of diagnosis
What that means	Newly diagnosed kids increasingly resemble the general population, not the high-risk population
What it doesn’t mean	Overdiagnosis, or that ADHD/ASD have gotten “less severe”
My take	This is the closest thing we have to a direct test of the broadening-criteria theory, and it holds up

What Did This Study Actually Measure?

Tarp’s team didn’t ask whether more people are being diagnosed with ADHD and autism. Everyone already agrees on that part. They asked a sharper question: has the type of person getting diagnosed changed?

Here’s the logic. Low birth weight is one of the most well-established risk factors for both ADHD and ASD. It shows up in study after study, decade after decade. If diagnosis rates are climbing because doctors are finally catching the true underlying rate of these conditions, you’d expect that risk-factor pattern to hold steady. High-risk kids should keep making up roughly the same share of new diagnoses, year over year.

If instead diagnosis rates are climbing because the net got wider (broader criteria, easier access to an evaluation), you’d expect the opposite. The newly diagnosed pool should start looking less like “kids with known risk factors” and more like “kids in general,” because the net is now catching people who would’ve been missed entirely under the old, narrower definition.

Using Denmark’s national registries, the researchers tracked 71,317 children diagnosed with ADHD or ASD between 2012 and 2022, compared against 713,170 undiagnosed peers, and checked how strongly a stack of known risk factors (low birth weight, preterm birth, socioeconomic disadvantage, family psychiatric history) predicted a diagnosis in the earliest years of the study versus the most recent ones.

The Number That Actually Moves Here: 54% to 17%

In 2012, a child born at low birth weight was 54% more likely to end up with an ADHD or ASD diagnosis than a child born at typical weight. That’s a strong signal — exactly what you’d expect from a well-documented risk factor doing its job.

By 2022, that same “advantage” had fallen to 17%.

Sit with that gap for a second. It’s not that low birth weight stopped mattering biologically — it didn’t. It’s that the population of kids getting diagnosed stopped being concentrated among the highest-risk kids and started spreading out into the general population. The researchers found the same attenuation pattern for preterm birth and socioeconomic disadvantage, and it hit ADHD diagnoses harder than ASD diagnoses.

Tarp put it plainly: “The key message is not that these risk factors are no longer important. What we found is that people diagnosed in recent years resemble the general population more closely than those who received the same diagnoses a decade ago.” That’s a real finding, and it’s a specific one, not a vague gesture at “more awareness.”

Why Broadening Criteria Isn’t the Same Thing as Overdiagnosis

This is the part that gets flattened every time this study gets summarized into a headline, so here it is as its own question.

1. Overdiagnosis would mean the condition isn’t real in the newly diagnosed group. This study doesn’t test symptom severity or clinical validity at all. It tests whether known biological and social risk factors predict diagnosis as strongly as they used to. They still predict it. Just less exclusively.
2. A shrinking risk-factor gap is what you’d expect from better identification, not looser standards. If assessment access has genuinely improved (more clinicians trained to spot ADHD, less reliance on a kid being disruptive enough to get noticed), you’d expect exactly this pattern: the diagnosed population starts looking more like the general population, because it’s no longer missing kids who don’t happen to carry a textbook risk factor.
3. Plana-Ripoll’s own read of the data is explicit on this point. “Our findings help us better understand why ADHD and autism diagnoses have increased so markedly in recent years,” he said, adding that the results “do not show that these conditions are being overdiagnosed or that they have become less severe. Rather, they indicate that the population receiving these diagnoses has changed over time.”

None of this proves zero overdiagnosis exists anywhere in the system — no single study proves a negative that clean. What it does is give the “everyone’s just getting labeled now” theory a real test, on real registry data, and the test doesn’t back it up as the main driver.

How This Fits With What We’ve Already Covered

This isn’t the first piece of the diagnosis-surge story we’ve written about, and it’s worth seeing it next to the others instead of on its own.

Adult ADHD diagnoses have surged fastest among young men, a group with almost no childhood infrastructure for the diagnosis. That lines up with a widening net rather than a shrinking one.

We’ve also covered the research showing ADHD is underdiagnosed at the population level, not overdiagnosed, whatever the dinner-table consensus says.

Then there’s the backlash to Channel 4’s “Great ADHD Myth?” documentary, which leaned hard on the same “diagnoses are exploding, therefore something’s wrong with the diagnosing” logic this Danish study directly tests and doesn’t support.

What TikTok’s ADHD comment sections actually reveal fits the same pattern: people mostly experience recognition, not self-diagnosis by algorithm, when they see their own experience described online.

This study doesn’t settle every argument in that pile. But it’s the first one built specifically to test the “criteria got looser, not the risk got higher” claim against a decade of national data, and the data agrees with the claim.

What This Means If You Were Diagnosed Late (or Mildly)

If you got your ADHD diagnosis at 28 instead of 8, or your evaluation didn’t turn up a dramatic case history (no expulsions, no obvious low-birth-weight chart note, nothing that reads like a textbook risk profile), this study is a specific, researched reason to stop treating that as a mark against your diagnosis’s legitimacy.

The whole point of the finding is that the diagnosed population has been moving toward people like you. Not because the bar dropped, but because the system finally started catching people who don’t fit the narrow profile it used to require before anyone took a second look.

A few ways to actually use this:

• If a relative or coworker throws the “everyone has ADHD now” line at you, you now have an actual study to point to instead of just getting defensive.
• Questioning your own diagnosis because you didn’t have a “severe enough” childhood story? That instinct is worth naming out loud to whoever diagnosed you, not something to sit with alone.
• Parents watching a kid get evaluated without an obvious risk-factor history should know that “no dramatic backstory” isn’t evidence against a diagnosis. Increasingly, it’s just what a diagnosis looks like now.
• Keep the 54% to 17% number in your back pocket. It holds up better in an argument than “more awareness” ever will.

Our Take

The headline version of this study is going to be “ADHD diagnoses aren’t fake, they’re just broader,” and that’s close enough to right. But the number that actually sticks with me is the 54-to-17 drop, because of how specific it is. This isn’t “doctors say awareness is up” as a soft explanation. It’s a measurable shift in who the diagnosed population is, tracked against a hard biological marker, across a decade of national records.

What I keep coming back to is Plana-Ripoll’s line that the findings don’t mean these conditions “have become less severe.” That’s the sentence worth screenshotting the next time someone implies your diagnosis is watered down because it came later, or easier, or without a dramatic backstory attached. The net got wider. That’s not the same as the net getting looser.

This is a Danish dataset, not a global one, and diagnostic practices vary by country in ways this study can’t speak to directly. But the mechanism it’s testing, broader access and identification pulling in people who’d have been missed under a narrower standard, isn’t unique to Denmark’s healthcare system. It’s the same mechanism showing up in the US diagnosis data we’ve covered elsewhere on this site, from the young-adult-men surge to the underdiagnosis research. One country ran the direct test. The pattern it found matches what’s been showing up everywhere else too.

---

This post cites findings from Tarp, M.E., Lousdal, M.L., Rask, C.U., et al., “Changes in Characteristics Associated With ADHD and ASD Diagnoses Over Time”, JAMA Psychiatry, 2026 (DOI: 10.1001/jamapsychiatry.2026.2722), and reporting from ISGlobal. This is not medical advice. Talk to a diagnostic specialist about evaluating yourself or a child for ADHD or autism.